Dear Mario, dear Jason, dear Linux folks,
Thank you for your great work on the Linux kernel.
On an MSI MS-7A37/B350M MORTAR (MS-7A37) even updating to their latest
system firmware 1.O7 released on their Web site on May, 23rd 2023 [1].
> Beschreibung:
> - AGESA ComboAm4v2PI 1.2.0.A. update.
> - TPM Out of Bounds Access security patch.
But Linux 6.4.-rc7 still disables the hwrng from AMD fTPM.
Jun 24 21:32:09.978772 tokeiihto kernel: DMI: Micro-Star
International Co., Ltd. MS-7A37/B350M MORTAR (MS-7A37), BIOS 1.O7 05/09/2023
[…]
Jun 24 21:32:10.259986 tokeiihto kernel: tpm tpm0: AMD fTPM
version 0x3005700000004 causes system stutter; hwrng disabled
Unfortunately, the warning does not say, what firmware version is
needed, and what the user can do about it.
The comment in the code, added in v6.3-rc2 with commit f1324bbc4011
(tpm: disable hwrng for fTPM on some AMD designs) [2], says:
/*
* Some AMD fTPM versions may cause stutter
* https://www.amd.com/en/support/kb/faq/pa-410
*
* Fixes are available in two series of fTPM firmware:
* 6.x.y.z series: 6.0.18.6 +
* 3.x.y.z series: 3.57.y.5 +
*/
Mapping 0x3005700000004 from the Linux warning to the comment, I assume
the board’s firmware is 3.57.0.4?
What AGESA ComboAm4v2PI would MSI need to ship, so the issue is fixed?
The previous system firmware version shipped 3.4e.0.4? (Is it decimal or
hexadecimal?)
[ 282.864564] tpm tpm0: AMD fTPM version 0x3004e00000004 causes
system stutter; hwrng disabled
Kind regards,
Paul
[1]: https://de.msi.com/Motherboard/B350M-MORTAR/support#bios
[2]:
https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=f1324bbc4011ed8aef3f4552210fc429bcd616da
On Sun, Jun 25, 2023 at 10:20:22AM +0200, Paul Menzel wrote:
> Thank you for your great work on the Linux kernel.
>
> On an MSI MS-7A37/B350M MORTAR (MS-7A37) even updating to their latest
> system firmware 1.O7 released on their Web site on May, 23rd 2023 [1].
>
> > Beschreibung:
> > - AGESA ComboAm4v2PI 1.2.0.A. update.
> > - TPM Out of Bounds Access security patch.
>
> But Linux 6.4.-rc7 still disables the hwrng from AMD fTPM.
>
> Jun 24 21:32:09.978772 tokeiihto kernel: DMI: Micro-Star
> International Co., Ltd. MS-7A37/B350M MORTAR (MS-7A37), BIOS 1.O7 05/09/2023
> […]
> Jun 24 21:32:10.259986 tokeiihto kernel: tpm tpm0: AMD fTPM
> version 0x3005700000004 causes system stutter; hwrng disabled
>
> Unfortunately, the warning does not say, what firmware version is
> needed, and what the user can do about it.
>
> The comment in the code, added in v6.3-rc2 with commit f1324bbc4011
> (tpm: disable hwrng for fTPM on some AMD designs) [2], says:
>
> /*
> * Some AMD fTPM versions may cause stutter
> * https://www.amd.com/en/support/kb/faq/pa-410
> *
> * Fixes are available in two series of fTPM firmware:
> * 6.x.y.z series: 6.0.18.6 +
> * 3.x.y.z series: 3.57.y.5 +
> */
>
> Mapping 0x3005700000004 from the Linux warning to the comment, I assume
> the board’s firmware is 3.57.0.4?
>
> What AGESA ComboAm4v2PI would MSI need to ship, so the issue is fixed?
> The previous system firmware version shipped 3.4e.0.4? (Is it decimal or
> hexadecimal?)
Sounds like this is a question for MSI - why they haven't updated from
3.57.y.4 to 3.57.y.5+. What can we possibly say about this from a
Linux perspective? The Linux code apparently is working as designed by
avoiding using the buggy firmware gracefully.